thermal and electrical resistance of metal contacts
TRANSCRIPT
Portland State University Portland State University
PDXScholar PDXScholar
Dissertations and Theses Dissertations and Theses
6-1-1967
Thermal and Electrical Resistance of Metal Contacts Thermal and Electrical Resistance of Metal Contacts
Roland E. Ott Portland State University
Follow this and additional works at: https://pdxscholar.library.pdx.edu/open_access_etds
Let us know how access to this document benefits you.
Recommended Citation Recommended Citation Ott, Roland E., "Thermal and Electrical Resistance of Metal Contacts" (1967). Dissertations and Theses. Paper 325. https://doi.org/10.15760/etd.325
This Thesis is brought to you for free and open access. It has been accepted for inclusion in Dissertations and Theses by an authorized administrator of PDXScholar. Please contact us if we can make this document more accessible: [email protected].
AN ABSTRA~ OF THE THESIS OF
__R_o_l_an_Cl._E_e_Ot_._t tor the . MS in Applied Science{Degree}
Date thesis is presented J_un__e_l~.9~·6~7 ~
Title Thermal and Electrical Resistance of Metal Contacts
Abstract approved _\~Ef Profess~
In engineering'praotice it is 1mpol~ant to know which factors atfectthe thermal and electricel res:i.s·tancee of metal contacts~. This thesisis to investigate some ot.thesetactors such as surface roughness and contaot pressure. Thermal ...electriaal contact resistance ratios tor met61lcontacts were calculated trom the experimental data. ..
The technical literature _was searched, and several papers were foundin whicn either ther.malor electrical contact resistance was studiedseparately.. However, none ot the papers recorded data for both the:rmaJ.and electrical resistances for the same samples. .
The information feund in these papers hae been used as a backgroundfor understanding the nature ot thermal and eleatr10al contact resistance.
Both of these contact resistanoes are primarily a function of theload on ·the contact and the condition of the surfaces. At low pressuresonly a small fraction otthe tota.lgross area -of the contacts is in metal ...to...metal contact. Inoreasing the load., tlattens the tthills tt and reducesboth "the ther.m.al and electrical contact resistance. This phenomenon iscalled ttspreading resistanoe;' since the flow of heat or electrical currentmust spread out after they pass through the restricted areas that areaatu.a.lly in contact"
Another type of thermal and electrical. resistance 1 Which is celledtt1ntertace resistance", is caused by a. film of foreign :material suoh asan oxide, etc. on the surfaces ot the contacting "hillett,
If'the space between the flhilla tl of a contact is fUled with air,there is a heat flow by convection currents. The literature indicatesthis quantity of heat flow is approXimately one thousandth of the totsJ.heat flow through metal contacts.
Since the only electrical current conduction mechanism aoting betweenareas not in a.ctual meteJ.lia contact is that d.ue to thermionic emission,the eleotrical resistance tor these areas will be extremely high at roomtemperature tor Which thermionic emission is negligible.
The experimentaJ. apparatus to measure both the thermal and electrioalcontaot resistances con.1sts mainly ot a bellows-actuated press which is
2
operated remotely under e. vacuum bell. The press. pressure loada the samplemetal wafers. A thin.rUm heat meter is used to indioate the quantity ofheat f'lowing 'through the metal. contacts. The temperature drop caused bythe oontacts is mea.f3ured with thermocouples. The temperature differenceand the quantity of heat flowing 113 used to calculate the thermal contactresistance. A strain ga.ge on. the bellows-press stem measures the loadingon the contact surfa.ces. Electrical probes are used to meaSU1"e the alec...trica1 resistance across the contact surfaces.
The thermocouples and eleatrica.l resistanoe probes are permanentlyinsteJ.led in the outer'two smooth copper wafers. This makes it possibleto qUickly change to other sets ot sample wafers of other metals andfinishes.
In order to use this permanetlt arrangement, it is necessary to .finishtwo uting surfaces ot the partioular set of metal wafers to be tested,s1Inilar to the permanent smooth copper wa.ters so that these two extramatins contact resistanoes can be found and thus be subtract~d from .theoverall contaot resistance. .
The data indicates that the thermal-electrical oontact resistanceratio can be changed by changing the load on the contacts.
The heat meter ha.$ pertorm.ed. very well, ~d this new method ofmeasuring heat flow W1ll undoubtedly become a standard method of measuringheat flux.
THERMAL Al'fD E,LEf:1'RlaAL MBISTANCEor METAL COITAcrS
by
ROLAND E It ()1"l
A THESIS
submitted to
PORTLAND STATE COLLEGE
in partial tu.1:rUlment ofthe requirements tor the
degree of
MASt'ER OF BCIDCE
June 1967
PORl-LAf\j 0 STATEo CO,LLJi,GEL.leRAR~
APPROVEDt
rread, Department or Applied Scienoe
Dean of Graduate Studies
iii
In preparation of this thesis, Dr. He.x-ry J .. White and Mr. Jamea W4
Combs of Portland State College and Mr il Paul Stangeland 0'1 the U. S.
Army Oorps of Engineers, Portland Distriot I gave me encouragement.
Mr. 1'1. E . Hager, Jr.. ot the Arxnst~ong Cork Com.P8.tlY proVided $.
sanwle heat meter which was very vital to the experiment.
Mr .M1ohatli Tierney of BU Electronics supplied. lIJample strain
gages and information on how to apply the gages;o
Mr. Jack Janacek, Head ot the Portland State College machine shop,.
has been extremely helpt'Ul in his efforts to put together the belloWf;J
press assembly.
My w1:t., Bethin(9, has encouraged me and. has been very patient all
during the ~evereJ. years I have been engaged in this graduate study ..
Cha.pter
1
2
3
4
5
TABLE OF CONTEIfl'S
Introduotion
Thermal Resistance of Metal Contaots
Electrical Resistance of Metal Contacts
Measurement of Therm.eJ. and Eleotrical Resistanoeot Metal Contaots
Conclusions
Recommendations
Bibliography
Appendix A ~ Description of Instruments
Appendix B .. Santple Calculations
Appendix C - Table 1, Summary of Ratios ofThermal to Eleotrical ContactResistance
Appendix D • Photographs of Apparatus
Appendix E • Graphs
iv
Page
1
3
7
14
20
22
23
24
26
OB:A.PrER 1
l:IfI'RODUa.FXON
THERMAL AND .ELECfRICAL RESISTANdEOF Mml'AL CONPAcrs
Nwne1."oU$ S;1:tuattQXl$ ari-., 1Q,eng1neer~gpractice 'in Whioh it is
~ortant to know the t});etrmal aud~ee~r1ua1 res1stancetJ,at 'the Juno
tionot two d.ry metal surfaoesi! Also it is helpful. to know how the
contact resistances $re affected by the cond.1t1on of the surfaoe, th.·
a;pp11Eid pr~$sure between. the surfaoes and. the tem,perature"
Electrical faUl1r$$ on clamped joints are due to hip contact
resistanoe. With sufficient electrioal oLlrrent floWing, the power
10'8 at th. eontaot may cause high te~eratures it eutt1aient heat is
not oarried awaY'. The ettect ot high temperattlre 1s to prodl.\o.$ oxida.
tion 'Which in. t\U'11 increas.. contact re.istQce~
fhe problE:$ ot thermal and eJ.eQtr1ea.l contact resl.t8J1ce it V$ry
~orte.nt in tb.e ope:re,t1on ot solid state deVioes Whiehare fabricated
'With dry contacts. These are jo1llts that are not we~cled 0):' $ol.dered
but Which mUlt dissipate beat ~ also 08:!!~1 an electrical current.
The pur,po$e ot t1l18 invest;t.ge.tion, 1s.to .hed some light on the
ratio. that are fiQStab11shed tor metal contacts under oert-.in surfaoe
conditions and mating prsssqree oan. be U$ed. to e$ta.b11sh the expected
thermeJ. eonduct1vity toX" a particular device by observing the eleo
trical contact resistance only. ~1$ would be an advantage since
electrical re$i$tances can be m.eaau:red more eaeUy than thermel res1s ...
tance in a oompleted devioe with metal surfaces in contact with each
other.
I searched the literature and found several pa.pers 1n wh1ch theX'
mea. oontaQt resistanoe or electrical oontact resistanoe e~er1.ments
were described. However, nona of theartioles inclUded measureDl$1t of
both the thermal and electrical contact resistanoes on the SaD1e .amplee;o
!fhe intonQation' found in these papers haa been uaed asa background. to
describe the thermsJ. and. electrical contact resistance prooesses. The
values that were reported 111 the art:Loles ha.s eeryedas So baai$ tor
estimating What data may be e~ected from 1.lf3 the,is e~rin1ent"
,In order to study both of the contact resistances simultaneously,
:I.t 1'188 necessary to load the contaot surfaces Ii' This was accompl:tshed
with a belloW$ press which used nitrogen gas pressure to load the metaJ.
~onte.Qt surtaoe$. A thermoeleotrio-type of th1n..foll hea.t Jl.leijer was
placed in ser1et;l With the metal waters to nteasure the heat flow acrosS
the contacts. Themoooqples indicated the temperature drop aoross the
.tal oontacts.. A strain Sage on the, stem of thE: press was U$ed to
indicate th~ loading between the oontacts. Potential type of inatr\!
menta we:r$ueed to measure the electrical resistar;l.ce between the eon
tacts. 'lhe 'Wbole assem.bly- was placed in th$ bell jar ot a vacu,um
sy8teru. 1norder to prevent thermal conve<:t1ol1 currentS.
ClIAPl'ER 2
'I'URMAL RESISTANCE' OFMmrAL CONtt'AcrS
Thf#re is a. clole analogy between-the thermal and electri.cal joint..
cond~ct8lJ,oe macllanisxn4 The concept of $PNading resistanoe d.eveloped
throUgh 8'tudy 01' eleetr1caJ.. contacts 11 adapted to the interpretatiQn
of the behavior of thermaJ. res1stanoepresented by a joint between
metal surtac$$ ina thermal circuit. (la, p. 259)
Data. tr~ the references has shoW that thermal contact res:18te.t1Ce
of joint$ 113 prilrJarily a tunqt10n of thQ load on the contact surfaces"
It is poslil1ble" at sufficiently high pressures or tor s\1fticientlr
smooth and flat surfaces tor a very large ;percentage of th.e surface
area ot $8.cb eon.taet faoe to be in·aetual metal ...to....m.et81 oontact w1th
the other. The thennal resistance ot the joint then 'Would be cons:t.d,.
ered to bea.pprox1me.tely equal to,.er~~ (l~, p. 259)
At low :pre~unu'es the areas are in metal..to-mete.l contact over only
a smell fraction ot .the total gross oontact area. The a.otual contaot
area is made up ot a large number ot·$epa.~te "hills tf or "point$lI dis ....
tributed randomly. !t 1s assumed that ita vacuum or even air tUle
the Hvoid" spaces between the aotual Qontact areas, the majority ot the
heat flows through the m.etal"",to..metal "hills" sinoe there is such lit.
high resistanoe to hea.t flow through the air path or a vacuum. The
thermal eonduatiVity through an air film is ~bout one thouse.ndtl1 of
4
that through the actaal metal. in oonta.ot II It should be m.entionecl..
however, that if hydrogen or helium gas. filled the "voidft spaces be
tween the act1.lSJ. contact areas, the oonveetion heattlow throUgh these
gases· would be approxima.teJ.y sj,¥ times that exPected throUgh an air
film_
'fhe resistance to heatf'low 1ll$inlyas$001e:tHit(1 with metal conta.ct"
can be discussed with the u.se of the concept ot spreadins res:l$tance.
In this ooncept,·the ttlines tf ot heat flow must "tan out" or I'spread
out" atterpe.ssins through the points of actual Qontaot II S1ncethese
contact ar$8S e.rerelat1veJ,.;y small conwared to the overall ~ea ot the
surface., the heat flux is very great at theJepointa. Also the heat
mu.st .nOw le.tterally at the surfMe to reach the contact points. As a.
load :Ls applied to the contacts, the projecting "h111slt start to tla:t
ten out and a larger total. area ;1$ available tor conduct iUg. heat. The
rougher the$urtacEz, the sreaterwUl be the effeot on tb,e thermal
contact resistance by incre$$1ng ordearea.sing the ::load" '!.'he presence
of an oxid.e tUm or other toreigtl material of low thermtil conduct1v1ty
also contributes to the thermal resistanc~ of the joint. liowever I ex...
cept at very low ;pttes-sures, the oxide resistanoe appears to aCQount
tor only $ s~part of the total thermaJ.. contact resistance. (12,
pp. 259*'260)
The oonclusions and resUlt$ ot three thermal contact resistance
ex,per1lllents obte.1ned from the li.t~r~ture are 1nolu(ied in this thes:t$
in order to establish some qualitative baeis tor indicating the ~or
tent fa.otors in the measuring ot thermal oontact resistance ~
Weills and Ryder came to the following conolusions based on their
experimental observations: (12, p .. 266)
1. The ther.rnal conductance of a dry joint increa.ses with pres
sure, linearly tor steel, and generally eXiPonent1ally tor a.1uminum and
bronze.
2;, The thermal resistance of dry joints decrea.ses with a decrease
1n the roughness of the surfaces.
3. At 8. given temperature, pressure, and roughness, the thermal
resistance of dry joints decreases in the order ot steel, bronze and
aluminum.
4. The thermal resistance of a dry joint decreases as the tem
perature 1ncrea$~$.
Jaoobs and Starr found the follolling results: (6, p. 141)
1. After studying copper, gold and silver contacts in a vacuWl1
they found that only in the case ot copper does the conductance vary
linearly with contact pressure. (1'he maximum pressure wa.s approx:1"""
mately 2.; kg/arn2• )
2. They found that the thermal conductance was approximately
the same when the surfaces were "just tou.ching lt or sepaJ."ated 'by a tew
mi11meters. This radiation conduction is less than 10.3 watt/cml°c.
3. In e.ll cases the thermal conductances are much smaller at the
lower temperatures ~ This is probably due to adsorbed surface films.
(The experiments were conducted at 25°0 and -195·0.)
6
Kouwenhoven and Potter experimented With steel samples and arrived
at the following oonclusions; (9, p. 520~5)
1 •. '!'hermalresiatance decrea.ses "With pressure in a. Il11\tU1er which
is essentially exponentia.l. The rate of decrease .is greater tor
rougher surfaces. For very smooth specimens the thermal resistance
i.8 pra.ctj,caJ.ly ind,€;lpendent of pressure.
2. The inconsistencies and insutt1oienoyof the data. make itim...
. possible to predict abe.clute values otthermal resistance for given
roughness andpxoessuras. It is probable that for heat flow xneaaure
ments the rm811 system of designating roughness is inadequate.
3.At constant press\.l.re and in tiheabsenae ot corrosiQn, thermal
resistance is .substantially constant as t~ll1Pera.ture leVel is increased.
4. There:1.s .need for more accura.te knowledge of the actual s~....
taoe areas in contaet as this remains one ·ot the greatest unknown
ta.ctors in the problem..
There is generally good agreement between the reported results of
the three groupe of thermal. contaot resistance experiments. One no"",
tlceable exc~t1onJ however, is the oonclusion by Weills and Ryder who
found that the thermal contact resistanoe of a steel joint decrease$
lin@arly with pressure (reported as thermal conductance of a dry joint
increases with pressure); whereas Kouwenhoven and Potter stated that
the thermal resistanoe decrea.sas essentially ex:ponentiaJ.ly for a steel
joint.
CHAPl'ER 3
ELEen-HICAL RESISTANCE OFME:rAL OONrAaL'S
Electrical contact resistance 1s a'very extensive sUbJect, in...
eluding a wide variety ot incidental phenomena, such as thermal et~
fectl:J1 coherer action and the behavior'ot different kinds of' surfaoe
fUms Ii A truly comprehensive treatment of electricaJ. contact phenom.
ena. would necessarily include the Palt1'er, Seebeck and. Thomson etfects
a.s well" HoweverJ theBe three effects belong to the d.omain of thermo...
electricity ratber than to e.lectr1cal conta.ct theory. Fora general
background, the review ot electrical contact resistanoe can be found
in an article by G. Windred. (13, p. 547)
The tour main factors that determine electrical contact resist ...
ance as follows;
1. The kind. of materials used for E!ach contact
2. The condition and roughness of the surfaces
3. The shape of the oontaateurf'aces
4. The mechanical pressure aoting between the contacts (13,
)),. 550)
In regard to materials, it has been found in general that mater!
aJ.S ot high electrical conductivity, such as silver and copper, give
lower electrical contact resistances than those materials of lower
electrioal oonductivity, such as iron and tungsten. (13, p .. 550)
8
A metal surface I even finished to optical flatness, baa many
points tha.t protrude" Thus, if two of these metaJ. surfaaes are bI'ought
together under very lov pressure, isolated. point~ on the surt'aoes will
touch. 'An eleotrical current will encoun.ter abigher resistance at
these restricted points. In addition to a. greatly increased ·current
density, "the cons'criatiol1 of the lines of current flow also causes$.
lengthening at these lines and thUS eauses what is mown a.s IIspreading
resistance", Hs ' between the contaots. If the oontact surfaces are not
perfectly clean, there iseUso an "interface resistanceft) Ri , caused by
oXide or ga.s films. The total contact resistance, Be == Hs + Ri " ( 8,
p. 458..8)
As the load on the conta.cts is increased, the "hUla" between the
contacts are deformed more and more untU the surface areas are m04."e
nearly mated. The maximum pressure that a given material can with
stand withou.t irreversibly deforming is based on the he.rdneliJs or yield
point in compression, H, measured. in psi. If ~ equ.a.ls the J.oad....
bear1ng~oontact area in square inches, we may write for ordinaq nat
surface contacts that Sb = 2F/H. (The ;force, F, is measlJ.reO. in
pOWlda.) Wh~never pla.stic deform.ation or $plintering of some of the
"hUla" in a. contact takes place, tllen add.itional subareas are "brought
into contact. When the force is applied. a. second ·time" this new oon,.
tact pattern will ¢ause a change in the electrical resistanoe. (8,
p. 458-s)
In addition to the path of electrioaJ. ourrent through the !lh1U."
ot a. contact, there is a.noth~r possible path of current flow tha.t
9
should be considered.. Thil! path is a.cross the IfgapJl between 'bhe con
tacts ~ This electric current would be carried by electrons that can
surmount the distance between the conta.ctSurtaces. However,at ambi
ent temperatures a sufficient number of electrons do not have enough
energy to carry any appreciable current aince ,the Itgap" between the
non..touching parts Of the surffl.ces1$ too great. Even, polished sur
faoes, flat to a qua.rtEitr wave length at light, still have "hills" and
"ValleY-fl" of over a hUIldred Rngstroll1 Units" R. (8, p. 458...s)
The high resistance of a. Ifgap't between metal surfs.ce$ has been
studied by J .. Frenkel. For a "gaptl of 10 R and a mean free path ot
electrons of length approximately equal to 100 R (whichrough].y corre~
sponds to the mean free pa.th of t:)leetrona at room temperature) the
resistance of the oontaet per unit surface area (in cm2) is of the
order ot 10..2 ohms i; For eo Tt gaptt twice as large, ,or 20 R, the resist.
ance is a.bout 108 ohms.. Further inoreases in the "gapn would mean
praot1cal.ly no: .flow of electrical current • For a "gap" of eo ~J it
would require an electrio field of 10 million volts per em across the
lI gap", corresponding to a potent:ts.l d.ifference of 2 volts, to obtain a
current of the order ot 10...8 a:aw/cm2• (4, pp" 1610, 1611)
Since the ltgaps" normally encountered :in metal. contacts are
greater than 100 R, any thermionic current between the contaots can
be neglected at a.m.bient temperatm:-es. Thus, it i$ assUIJ1ed that at am.
bient temperatures all measurable electrioal ourrent flows through the
actual contac.t /thUle" and the film on these "hUle tf• (At high
10
temperatures thermionic .....:1.on of eleetronswoUld carry part ot the
electr10al aurrtitnt ~ )
The resUlts -.a conclusions of three $leetrieal oontact resistance
eucper1ments obta:S.nQd trom the li1;eratn,1r~ are 1n.Cluded 1n this thesis in
order to ."tabl11Jh a qual1tat1ve ba,!. tor1nd1cating th.e important
tactors1n the u$Suring of electriQa.1 eontact rea1etano$,.
Ko\tWenhoven and Little found the tollo"ins results t ( aI P II 464~s)
1. '1'h~ valu.e of sPnJetdinS resi$tauce can be effectively eon.
troll.a. by varying the number ot eontaC'teUbareaS, linn It They found
tor th_ mater:i.al8 tested. (sUyer i alUlIdnum, bronze and brass) that tor
un'· eq1.JJJ, to about 3000 ( ..pproximatel11~hOOO tor .. contact of one ,
$q,uare :lnCh 111e.r$e.) the valu$ Of Bs becomes l~.s than one m.1Qro~ObJn,
and. efteottvely d1$$ippe~sJ being n,egl1g1ble compared to R!. The load
on th~ one·-halt inch d.1am.eter speeintens R$ 230 pounds oX' mo~.
2. Fo~ thin aurfaeefu.m., the :J..ntertaee resistance" Rl , may be
controlled 1:>t~ the load on theoont$.ots.. Hi is independent of
the cont.ct pett$rn*
3. Increasing the foroe. on the contacts will reduce both the
epreadtDg and interface r.8i8t~ae.
G. W1M:red report~ the re,u1te ot Dr. Eberl:1ard Cont1U$' 8 d1aser..
tation on the 1nf'.luenC$ 0'1 the _gnitude ot pressure and liI~t8C(t u,pon
electrical contact re$1stance. ('lbe oX"iS1neJ. ;pe.pe~ 'WaS Published. in
Dresden in 1929.) SOUle of the eonelU$ions XDade by Dr. Cont1us are
sUJlQ118.rtzed as tollolfe: (l3, pp. ;1'7, 519, 581)
l!l In the case ot pointed oontfi.ota the p,res8ttre and re81stal).ce
are related in aooord.e.nQ~ with the formula R= k/{p)1/2. The actual
or ettectlveeontaot surtace approximates elose1y to t1l.ft apparent
surface.
resistance. It.. however, the surfacEte are very rough, the oonditions
become $1mUar to those of pointed. contacts.
4. In the cU, of ground end stnooth .urtaoes the Qont$.cts ot
.1."1"88 ~ea have a high_X" rel1etance than the smaller on.s. In the
casca of normslly t1n:iahed. surtacfl!$ the:re 18 no appreciable ditte:rence
in thi. r$$peot.
A portion of Table I, rtftelat1on. Between Contact Rea1.tance arJ.(.\
Pre••ure. (Ell C:ont1u8) PI :
Condition$
12
VeJ.ues of n endk in Resistance
Formula.
R =k/p4n k
Thfloretlca.lEmp1r:1.calNOrlll$lf:1n1.hGroundj%lewGround, newGround" newGround, m,uch used
1/a1/212225/3
0.0002040.0002300.0004300.0100000.0500000.4009Op0.001900
€h E. Luktl reported "the following: (10,. p. 68)
1. With te$peet to temperature, the contact resistance of the
metal. var1edJ tl1'U.4l,.ome tests W$re made 1dth copper in 'Whioh the
contact re$1..tance was slightly decreased with ~cre$s1ni t.~erQture8.
~h1$ would 1n.d:1ef:\te a nept1ve temperature coetf1eleJ',1.t. lJhis l4$S ver.
1t1ed by tests on the resistance of Qopper oxi.de, which had a spec1t:Lc
resi$tance of 39,400 oh:rna per 1nch cube atz.ro t1ep-ees C and 2160 obn1s
per lncb. cUbe at 3.00 degrees o. The.oontact :res1$ta.na. at braes and
2. Based on a large nUlrlber of te$ts on large copper bU8 ...bars, it
has been stated th-.t the eontaetreaistancEJ is inversely proportional
to the p~tl$a\lrth'1'husI BP == 0.0008 to o.00J,,6 tor a cleaned copper
sUX'tace Whe~e R 1.$ the coX),tact r~.i8tanoe in ohma and P is the pre$~
sure in pointEl, This equation says that the res1$tance of such a eon
te.ot 1$ independeat 0'1 the area of contact &ld Q.epends only ~on the
total p:ressure applied. Tb,ua, tor a pressureot 1000 10$ .. per s.quare
inch, the contact :r1esiatanee given by LUke 1$ 0.9 miero~ohmS per
13
square. ineh. Expe:r1mentaJ.~esult$,ind.icate that bras, andirQ%i do not
foJJ..QW J~Uqb.e. simple equat1on.
'l'he.ri .1$g@$l."al:ly good agreement between the re.u1t$ ot the
three groups of f!lectr1eal contact resistanoe $.X.per1ments reported by
the nIHil.rollers.
Tbe formula R .. k/p:1 given by E. Oont:1ws tor~ormaJ. tintshed
plan.e aontaet$ aom.eWhat resemblea the to~uJ.e. .J:lP == 0.0008 to 0.0016
~iven by G. E. Luke.
C!APrEa 4
MmA,STJREMEI'f OF '1'HEaMAL Am> ELECTltrOALRE8:ISTAWCE OF MmrAL COB!'ACTS
Measurement elf the thermaJ. and el~otr:i.aal resistance$ of metal
oontacts" 'Was aecong;>11ahedWithan apparatus that wae operated ~ a
vaau.Uil. This arrangement 'Was used in orde,r to prevent heat floW' away
trom th~ $~ples by air cottveetion currents.
A bellows piress was used. to load thEt oontactaurfaces. ~hU$, thEt
load on the t8lfq)la metal watersoould be changed b'OXD. outside ot the
vaouum bell by _rely ¢hanging then!trogen gas pressure in the bel·
lows.
The; bellows press con.i,toe of a. base With tour .tee! rod.$ which
support a steel upper plate to which thE1!bellowtI itJ, attaQhed.. The
~~port rods have a lons'tbread. on the upper end in" o:rQ.er 'ho 'permit
adjustm,eliltot the height ot the beUow assembly. A iJteel laMing
stem i_conneoted to the bottom Q'I the bellowa aQd is 1\114. by two
plates' mounted on the Sl\'P'port rodS.
A platen is engaged on the end ot the loe.din8 stem with 9. bsll
jobt in o:"der to :tnsure that the contacts\1rtaaes Will "be loaded
eVQn1yIt Resistance Wire is wound around a $lot which 1$ aut .iD. the
upper platen to provide e. liJo\U"ce ot h~atlt The lower plate was used
to oonduct the heat away.
A heat meter, Which consists ote. 'thin ribbon of $emiconduato~
materlal1 1s mounted 1naer!e$ with the metal wafers. ':fht meter
15
measures the heat floW' through the contact surfaces ot the metal
waters. (A mored.ete,Ued desoription of the heat meter will besiven
late:rll) 'I'he s~le :m.&taJ. wafers art! plaQed between the heat meter on
the 'bottom and the upper permanent waterOll the toPOl
.The thermocouples and electrical resistance. probes are permanently
installed in the outer two smooth copper waters,,· Thie msJfes it pOQ$i
ble to qUickly ohange to other sets of· wa.ters ot o-ther metals and
t1n1ah~s.
With the data ava-Uable frolU the thermal and electrical resistanoe
tests on the e1ng1e smootl1 copper surfaoe, it W8.8 possible to calculate
the th~rntal .a.nd. e,J..EH:tr:Leel resistanoes ot other metal contacts .'1'0
acconwl1$h thi$; two water$ ot each of the br$.ss Ilnd steel meta.l$wEtre
finished all on.e side e1mUarly to theln100th copper' finish and then
copper. plated. Thi$ perJJ1itted the sets of the other metal waters to
be mated. 'd.-lib. the permanently bstalled $mooth copper waters. Sinoe
th<t theX'J'J'Jal a.nd alect1"'1aal resistances aOJ;'QSS these two outer smooth
copper oontacts had. be~ calculated, 'the net contaet resistances for
the t.st samples could. be determined.
'l'he load on tbe metal waters was measured nth a BLi Eleotronios
stra1X1 i9.gEJ cemented to the stf.teJ.. load1n6 stem of the bellows pre$s"
The aotualmeasurem.ent of the heat flow quantity is one of the
sreate.t problems 1n $ thermal contaot resistanoe e~er1ment~ The
teJl\P$rature~gradient In$thad hss been used extensively in previows
experimentsS however I th:1s Illethod ot ~asl.lring heat fl()W haB $ poss1ble
source 0'1 er~or.. O1U.y a snaU error in the measurement ot the $lope
of the gre41ent can causes. large Change in the apparent h~at flow1
espeoially it the toteJ. temperature difference across the resistanoe
iEl J1.ot very great.
In orde)." to insure $. more a.courate measurement of heat flow, I
selected a. Iiewmethod·ot determ1ning htat flow across the mettU con..
t~ots with th$ use at a heat meter. 'rhis new way to measure heat now
waa developed by Mr. N. E. Hager; Jr. ot the Arn1$trong Cork Company,
The heat is measured With a "Thin·Foil Heat Meter".
!:Chs hElat m~ter used in thie ex,periment eonsi,ts ot a. thin....f'oil
differential th~rmocouplewhiah 111e8;$U,re$ the te~ere.tl.lre differ~oe
between the two face. of $ thin electrical insulator II It i$· an abso
lute 1.nstrument; i. Q II' its calibration t.ctor can be oalibrated &$ a
function of t~erature trom th~ phlSicaJ, and geometrical parameters
ot the tUm and the thermocouple material., and :0.0 (Jostly experinlental
calibration is needed. 'J1he pa.rt1QUlar heat :meter used 1$ rated within
! 5" accuracy. The devlae is toundtosettle to a stable reading in
less than Ol!5 $ec.The xo.eterha$ negl1gibletherm.al res1stanoe, and
its pre,ence does not upset the thermal field under study.. Thes$$d""
vt\11tases are obtained at the expense ot obtai.tling a very small .em.t
With normal environmental heat fluxes • But the electrical resistanoe
of the meter 1$80 low that good readings aanbeobtained With modern
mierovoltmeters.. Heat t,luxes as lQW as 0.05 Watts/me have been de.
tected with 2 seoond response. (5, p. 156,)
The ll1$tal water$ are 1/8 1n0h thick (0.317' em) and 9/16 inCh in
die:m.ete:r (ll! 4287; om). The wafers were made this partieular d1amQt~r
17
so that their area would be 'equ.al to 1.11 6 amP.. The area. ot 1.6 cm2 WEl.$
used for oon-tact re$istance exp$r:lment$ wh10hwere conducted by E.
Contiua.. He developed aome xoelati.Qllship$ between f<tlectrieeJ. contact
res:l.stanoe and the loading between metal contactslt
'I'hree seta ot waters weretiniahed. t.O 81mooth surfaoe 'by ti~~t
using a #400 fine grade' $andpaper and then polishing' them with Crocus
cloth. The other three sets ot waters were finished with e. No. 80,
rough emery cloth. Care was taken to keep th$ surfaces flat.
The 0.25 1ncb d.iameter steel stem permitted a convenient rela.
t1onsh1p between the kUogre.ms load. and the micro..,inch·ind1oation on
thE! meter. The strain sage read 7 t,5 mioro"inch increments toX' changes
ot ; kUogreuns in load.
~b.e vacuum. i!l1Stent oons1sted ot a mechanical vacuum pump and. d1$'
oharge chamber. Since a. very high. V'$ccu,um was not requ1red, the diffu·
sion PWll.P apparatus wa£I. not used. A vacuum was maintained during the
tests $.1; a.pproxtmately 0.5 mm ot Hg,
The electrioal potential was mea.sured with a current ot 0.035 amp
tlow1ngaoross other metal contacts. This current also produ.ced a
voltage drop in the leads as well as in the sold.ered joints.. It we.$
necessary1 therefore, to determine the voltage drop in the entire ex....If
tfiarnal cirou:l.t in order to isolate this resistance from the contact
resistance.. This loss was mesaured. by soldering a wire between the
leads to the permanent copper wafers whicbaated a.s the electrioal..
potentisJ.p~o'bes. These two soldered connections e1mulated the two
soldered eonnectiOD.$ to tlle oopper water probes.. The wire was
18
measured to equal the lead lenstha between tb.e.se soldered connections
and the teX'lldJ:lal pointe at the copper water$,
'I'wenty-..t1ved1fterent read1qswer$ taken a.t 0.035 ma, Q..o40 ttI8,
and 0.04,.. fhe resistor was changed each time to control tb,e eur
rent II
An :lnterest:J..ng thing occurred in the px-oceas of msasurins the
current.. ThemUl1atnp xaeter had a bent pointer1 but it 'WaS the 01\1.1
one availaPl_ to me.. A thought ooeurteCl to me, however, that this
bent pointer might be USed to a.dvanta.ge~ I reasoned that pOls1bly the
bent ·pointer coul.d. be used as a vernier,' Th.U did oocur since the bent
pointer S:r.8duall.y covered. a meter' 1n4ic8ttlng Ill$rk as 1t moved very
s11ghtly.Xt gave the &a!pearanoe of a Qb.ans1ng length tromvary Ihort
to .. long black line. With thi. arrangement I W8.$ able to reproduce
readings ot the extenlal resistance circUit with the 50 ma full scale
meter to W1th~ trwo.tentbB of a m:1Qrovolt plus or minus of the $ver~
of allot the rtlad1ng$ taken at 0.035 lQlh ~he potentiometer read to
tenth$ of a miorovolt ..
Th~ eleotr1c$l resistance re$.d,1ngs of the .mooth and rough steel
samples we:re taken Vtth no beat grad1e.nt since the thermoe.l.eotrio E!t.
teet of the copper.$tee~ oouple would. ··aAd a cons1dera.b~e potentia.l to
the circuit. The eleotrical rasiate.nces Qf the brass $aJt;)lee were
taken nth a. heat gradient1 but this effeot is very minor since there
18 $uch a small tbermoelectric potential dittar~nc$ between copper and
bra$$ • This oould 'be considered $8 asD1$.ll sOt.U'ce ot erroX' for 8.n:Y
:ruture ex.per1mentsl work, however.
!rhe u:t:timate goaJ. of my experiJnental work was to determtne the
ratio of the theX'lD$J. to electrical. oonta4Jt resistance at progre$'1.ve.1.y
grea.ter 1.04\d1n.g 0'1 the oontacts. The S~1ot these cateuJ..at10nsis
g1ven in Tab.1e 1, Appencix C.
CHl\PlER 5
CONCLUSIONS
1. The ra.tio of thermal resistance to electrical resistance can
be ~hanged bychariging the force between metal. conta.cts ..
2. The thermal resistance decreases more than. the electrical
resistances for the sam~ increase in loading.
3. There is e. greater change in the ratio of thermal resistance
to electrical re6istance tor smoo·th copper and smooth brass contacta
for a change of load. thad for rough copper and rough brass contacts
tor the same change in loading.
4. The electrical resistance of' the rough steel contacts was
measured to be very roueh less than that tor the smooth steel contacts
tor the same loading.
5. The apparent increase of the Rrr/RE ratio for rough steel at
higher loading 1$ based on the large decrease in the measured eleotri..
cal resistance since the thermal resistance continued to decrease at
the higher loading"
6. The statement by Kouwenhoven and Potter that the tao"1; that
the rate of decrease of thermal resistance for steel with pressure is
greater for rougher surfaces is "born out If by Fig. 2 of this thesi$"
(9, p. 520~a)
7" The heat meter has performed very well, and this new method
of measuring heat flow will undoubtedly become a standard method ot
measuring h~at flux.
RECOMMENDATIONS
1. If helium gas could be used to surround the metal. contacts"
the increase in the thermal conductance between the contacts could be
m.easured. Also, tests could be run in air at atmospheric pressure.
These two tests could be compared with the tests of the contacts in a
vacuum. The test in helium should give a gas conductance approxi..
mately six times as great as with air between the conta.cts.. The teste
in air could be used to check the value, given by Weills and Ryder,
which is about ;one-thouse.nd.th ot the total heat that flows through the
metal contact.
2.. Tests coUld also be conducted VIith carbon waters if one side
of two of the carbon waters 'Were covered with a copper fUm depo$lt.
This test would indicate the magnitude of non-metal contact resistance.
3. Tests at very lOll pressures and a.t higher pressures would
give useful data.
BIBLIOGRAPIIY
1. Brunot, A. W. and Florenoe F. Buakland.IJJhermaJ. conta.ct~ J,·esistMee of laminated and machined joints. American socie'ty 0'1meohanical engineers transaotionG" 71: 253-257 • 1949.
2. Flom, D. G. Contact resistance measurements at low loads" 'l'hereview of scientific instruments. 2.9:979....981. 19;8.
3. Flom, D. Gc, and R. H. Savage.General electric re~iew.
Insulating fUms on metal contacts.59·6~ 1955.
4. Frenkel, J. On the electrical resistance ot contacts betweensolid conductors. Physical review. 36:1604-161.3" 1930.
5" Hager, NathanieJ. E. i Jr" Thin foil heat meter.; The review ofscientific instruments. 30:1;64-.1570. 1965.
6'. Ja.cobs 1 R. B. and C, Starr. 'l'hermaJ. conductanceot metallic oon-tacts" The review of soien'hific instruments" 10:140.141.1939.
7" Kisliuk, P, Using contact resistance to measu.re adsorption ofgases on metals. The Bell System technical journal.. 924.949. 1958.
8. Kouwenhoven, VI. B. and Cla.rence W. Little. Contact resistance.The welding journal. 31:457-$·.464.$. 1952.
9.. Kouwenhoven, vI. B. and J. H. Pot-ter" Thermal resistance ot meteJ.contact$ 10 The welding journal. ~(: 515"s."520-8.1948.
10" Luke1 George E. The resistance ot electrical eonnect1on$.. Theelectric journaJ." 2l: 66.69" 1924.
11. Powell, Robert L. and Arthur A. Aboud. Electrioal contact resist..anee ot copper"'copper junctions at low tempera.tures. Thereview of' scientific instruments. 29:248"'9. 1958.
12. WaLlls, N.. D. and. E. A. Ryder. Thermal resistance measurementsot joints formed between stationary metal surfaoes.. Ameri..can society of mechanical engineers transactions. 71:259267. 1949.
13.. Windred" G. Electrical contact resistance. Journal of theFranklin Institute. 231: 547-,85.
APPENDIX A
DESCRIPrION OF INS'l'R'UMEIf.rS
Potent:10IQ.ete~
L$ediJ and Northrup Company, t~e K-Z, No. 1133030, used withe.standard. Qell,1.0185 volt., range .1. x 10-7 to ;L.6 Vo.lts~
QalvanoUietel:'The Ea1.1ng (Jorp., ffSoalempft, Oat. 1o. 7891/08, 2;oh.nl$.
StJ:'a1n. .Indioator (For s.a.4 strain gages)Bs.ldw1n~LiJna-ia.milton,Type N,measures in micJ.'"o..1nches.
DC l4ilJ.i!!RMeteX"Sene:s.tlV$ Reeearch Instrument Oo:rporat1on, Model tfUnlvEUI;ilJit7U~
No;, 921,48, range Q,,-;O ma.
Heat Met$rArmstrong Oork CompanYJ No. 61.6664>.. th111....toU type (&:rea,
1.44 am2).
VaouWll (]ELse. Indicator and. ProbeItOR Equipment Corporation, type 701.
APPENDIX B
SAMPLE CALCULATIONS
At 10 Kg lo~ tor smooth copper waters
Temp. Dift'.' q: (239. O...198~ 5)lO~6 volts43 .. 217x.lo ...6 volts per 00
Q == Heat Flow == 2l,2xlO·6 volts x 1.44 emS (we:tta), 1540xlO·6 volts per ami
cm2""OCRlrc == ThermaJ. contact res1$tanae, watt' tor one joint of oopper
K .:If Q/Te1JJ.fi" Ditt..
RTc == Thermal resistanoe of' six 118ft thick Qopper waters
25
R ::;: TotaJ. measured el$atr1caJ. resistance was measured at 0.035"emp
~ =Electrical rei1stance of leads to permanent copper waterprobes =625.0XlO·6 volts ,
REe == Electrical re,sistanee of six 1/8" thick copper waters
# (1.7&10...6)(0.75)(2,54) == 3.4xlO~6 ObJn/cm2
At 10 Kg load tor ,rough coppu wafers
Two mating surfaces, ot the assembly 'of the six rough. finishedcopper wafers, were tinished simUarly to the smooth copper wafers;therefore, two of the res1.tances, caJ.culated tor smooth copper contaots, were' elso sUbtracted from the total oalaulated resistances torthe rough oopper waterS.. This left flva rough surfaoe contacts.
REO = ((638:g3~25'O) 1.6ltlO...6 ~ 3.4xlo~6 ... 2x74.2d.O~6) t =42.8xJ.o·6
Note: The values for ttL'· wers kept in numerical units only toroomparison purposes.
APPENDIX C
TABLE 1
SUMMARY OF RATIOS OF TBEaMALTO ELECTRICAL CONTACT RESISTANCE
26
KgLoad
o
5
10
15
20
25
30
35
40
45
SmoothC?;wer2.84
1.46
1:.29
1.07
0·99
0.94
0.74.
0·75
0.66
Rough SmoothCopper Brass
2.10 3.78
1.86 2.14
1.66 2.50
1.63 2.04
1.34 1.95
1.17 1.84
1.10 1 .. 61
0.79 1.54
0.88 1.41
0.81 11155
RoughBrass
3.34
2.62
2.41
2~16
1·75
1·55
1·50
1.50
1·32
1·31
Smooth RoughSteel 'Steel
0.131 ***0.110 1.0
0.124 0.87
0.117 0.80
0.161 0.69
0.192 0.72
0.196 0.80
0.178 0.94
** 1.59*
** 13.38*
* These ra.tios were calculated on an observed very loW' electricalresistance.
** Failed to get a heat meter readins.
*** Failed to get an elect~ical resistance reading.
Note: The above ratios a.re xl05•
-=- ~~r-•• ;- '4 ~
APPENDIX D
MAY 67
Equipment assembly of experimental apparatus
MAY 67
Vacuum gage, below table) galvanometer andpotentiometer on table, and vacuum ring andbellows press in the background
27
MAY 67
Nitrogen supply pipe to vacuum ring and bellowspress) terminal connection box and the strainindicator on the right
MAY 67
Nitrogen pressure gage and relief valve) vacuumring) bellows press and platen heater whichpresses on the wafers
28
,.-,-,.----
APPENDIX E 29
...
Fi9'l.i f~.I'
:: = :Wj!- ~t
= . ~"
- lJl-- .~·t=--::::::=====--
--,
-
--~' 1
'-
-
-
-
...
/
'.:.J=-
.11- ..-
:. 1--1-1-1_
:::
-
iII·
.... ... ~ .
r-" !:"A', ,
,I ....._. 7
ii'_
1/
'"'~r~ ~
7 "- )·If.iJ.',~~t
~ 1/
too- ~.jj~
6 '"/, 8' ,10I ',j ,II' ',i, ,:1I,
330
2.5
", :"
1.5/5
, ".. ' .
7,,;.8 . 9 10.','., '11)
2:2.0
.,' :"',' I-odd- K,')tJlrClms. ~. . .. ", .
, "
,04 5. 6
\ '
'S-
.."
II)J
f-++""f-t+f-++H u1+.j
j
t
F/g.. 2
30 '
t, ,. I
,I1
..... ....."
.... ...... ~
~ ~..... ,..,
....... ~
"'" '4,.
'-.J ~
"-...... "'"... ~il
....~
~~~~~lr
+.
'4 1.5
/S2 2.5 3
20 :fGJ.. 5
40 S{)6
'·1 ;!'I
31' . :
'0,IJ ~.rJr
. 11:1.
t
6 ,1 8.. 5
-10 5(J2.5 3
.30
IL
2
201.5
/.5
....... ., , .. "~; .11":1
l.Cl\,o-a r I [;j 1.
, I
L\ 11 I l;:; .... I?
l-
... L()(1d-- Ki/o.frdms
:'.' ..:, a1iOR11AND 1YATICOlLEGE LIBRARY'," .